Google Analytics should be everyone's first web analytics solution
Updated January 11, 2019
Google Analytics should be everyone's first web analytics solution
Score 8 out of 10
Vetted Review
Verified User
Software Version
Google Analytics
Overall Satisfaction with Google Analytics
Our business is an online website. We use Google Analytics to track site traffic, overall engagement on the site, key events that happen -- all with an aim toward understanding our user base and what is working and not working on our site and our efforts to increase traffic and engagement (and monetization!)
- Aggregates statistics along multiple dimensions for quick analysis -- geography, traffic source, campaigns, landing pages, etc.
- With goal tracking, it allows you to see conversion rate metrics (be they site registration, purchase, or whatever) across various landing pages or audience segments
- It's very simple to create sets of queries and group them together in dashboards of key metrics and trends and share them with colleagues across the organization
- Google Analytics does allow you to filter results based on a primary or secondary dimension (e.g. landing page and source) -- but it would be better if the user could pick finer grained segments utilizing more dimensions (e.g. "visitors from Canada who visited Landing Page X via Inbound Referral Link)
- Comparing results over time (week on week, etc) should have shortcuts for common analyses (vs same day last week, vs same month last year, etc)
- With Google Analytics (and all their products really), Google will change things by adding functionality or changing the names of concepts (recently they changed "Visits" to "Sessions") without so much as a peep alerting the user base. While most of the changes are for the better, a heads-up would be professional and appreciated.
- We spend far less time assembling quick views of web data for ad hoc analysis than we otherwise would.
- We are able to track website conversions at a fairly high granularity with far less effort than we would otherwise require
- Because it's free we can give access to Google Analytics to all our employees and bring more transparency to our operational performance and enable the curious to discover new insights.
KISSmetrics is best suited for instrumenting specific conversion funnels and looking at individual user behaviors. This type of by-user analysis is impossible with Google Analytics, which is understandable given the amount of data storage that would be required (for a free product!). Ultimately the tradeoff for us was that aggregate statistics along multiple dimensions was 'good enough', though we may revisit KISSmetrics [or Mixpanel or another similar solution] in the future. I don't see them as mutually exclusive. But given our time constraints and desire to deal with only one solution, we went with Google Analytics.
Using Google Analytics
6 - Operations, Business Development, Finance, Development, ... basically the entire company uses (or at least has access to) Google Analytics.
1 - Really you just need :
- Web development help to get the tracking snippet embedded in the right page(s) on your site (and embed any event tracking you want to do)
- An analyst or someone with good analysis skills to setup a set of key views and dashboards for measuring performance
- Analyze web traffic (visits, pageviews) by source
- Analyze site engagement (pageviews, time on site, bounce rate, etc..) by source and page type
- Analyze conversion rates (site registrations, purchases, etc) by source and page type
- We developed a widget that allows other sites to embed pieces our content (as a teaser to click through to our site). By setting up a separate Google Analytics instance for exclusive use by the widget, we can isolate views and clickthrough rates on the widget without having to filter them out of our other site traffic (which we would have to do by default everywhere otherwise). So my suggestion is make use of the ability to setup multiple Google Analytics instances where appropriate.
- By using non-interaction events, we are able to measure engagement and time on pages that would otherwise count as bounces with zero duration. This is important for separating the effect of a true bounce (this page wasn't useful to the user and they immediately left) from a softer bounce (they came and read the page for 5 minutes but then didn't see anything else interesting and so left the site). We set a threshold internally and ping the GA server every 30 seconds or so but only send interaction events after a minute to ensure the person was really engaged before we "debounce" them.
- With Universal Analytics, we're able to send analytics events from the server side of our application to the same Google Analytics instance, which can be very useful in avoiding client-side race conditions and other threading issues. UA can also be used for tracking across not just server (vs client/browser) but mobile apps as well. If you haven't upgraded your Google Analytics instance to Universal, you should.
- We plan to make use of the audience segmentation features (to differentiate between behavior of registered vs non-registered users).
- We plan to make broader use of the Content Segmentation features which are quite handy when you have a site like ours that is dynamically generated and contains tens of thousands of pages (and growing).
Evaluating Google Analytics and Competitors
- Price
- Product Features
- Product Usability
- Prior Experience with the Product
- Vendor Reputation
Since we used Google Analytics from day one as a startup, the fact that it was free obviously played a major role. But beyond that, I had familiarity with Google Analytics from a prior company / role, and knew that as we looked for SEO and other consulting help, that virtually anyone who knows anything about web analytics knows how to use Google Analytics. The ecosystem of practitioners is strong. So even if you "graduate" to something else later, Google Analytics is always a fine place to start.
I would not have spent as much time looking at KISSmetrics honestly. We may revisit it as an add-on for deeper dives into individual user behaviors but Google Analytics gives you a lot of runway for instrumenting, analyzing, and improving a site (or app) well before you need to think about replacing or adding to it.
Configuring Google Analytics
If you've got a large or complex site that's dynamically generated, make use of the content groupings dimensions which will make it far easier to track behavior.
Instrumentation using custom dimensions is also recommended (for user IDs or other attributes about their user or their session or even experimental page treatments).
Instrumentation using custom dimensions is also recommended (for user IDs or other attributes about their user or their session or even experimental page treatments).
Some - we have done small customizations to the interface - You can setup custom dashboards, but they're really only good for very high level directional views.
Once you drill into an analysis view, there's very little to zero UI customization -- you can just enter query parameters etc.
Once you drill into an analysis view, there's very little to zero UI customization -- you can just enter query parameters etc.
No - the product does not support adding custom code
We built a wrapper around the client-side code that fires google analytics events to also fire them to other systems (like Pendo for product adoption tracking, Optimizely for A/B testing segmentation, etc), but no custom code within Google Analytics itself.
Using Google Analytics
| Pros | Cons |
|---|---|
Like to use Easy to use Technical support not required Consistent Quick to learn Convenient Feel confident using | Lots to learn |
- Analyzing traffic and its relative performance along multiple dimensions
- Setting up dashboards of commonly reused analyses is quite easy
- Analyzing traffic along two dimensions (and possibly filtering on either or both) is fairly cumbersome and requires too many clicks. For instance, looking at traffic by landing page type (a content grouping) and then narrowing down into traffic that came via SEO requires a bunch of repetitive steps (vs there being a shortcut dropdown for traffic source type as there is now for content grouping... )
Yes - I've not installed any dedicated Google Analytics mobile interface, but the standard web interface works pretty well on an iPad. It is not a good experience on a smaller mobile device (e.g. iPhone) -- be prepared to use the pinch and stretch a lot.
Upgrading Google Analytics
Yes - We upgraded from the older Google Analytics to the Universal Analytics.
There was a completely unexpected impact in that the event logging mechanism changed but the change was not clearly documented and the old code failed silently on the new platform. Specifically, marking events "non-interaction" uses a different semantic in the newer Universal Analytics version, and if you use the old syntax in your event logging, it not only won't work, it will fail silently. That's a bad design.
There was a completely unexpected impact in that the event logging mechanism changed but the change was not clearly documented and the old code failed silently on the new platform. Specifically, marking events "non-interaction" uses a different semantic in the newer Universal Analytics version, and if you use the old syntax in your event logging, it not only won't work, it will fail silently. That's a bad design.
- Universal allows us to log events from the server side of our application more cleanly (and mobile in the future)
- Since they tend not to announce new features or interface changes before just springing them on the user base, who knows?